Note: This is a fork of the original PocketFlow by Zachary Huang and The Pocket. All credit for the core framework goes to them.

This fork adds extra utilities, including a Universal LLM API Client for easy access to 8+ LLM providers.

English | 中文 | Español | 日本語 | Deutsch | Русский | Português | Français | 한국어

Pocket Flow is a 100-line minimalist LLM framework

• Lightweight: Just 100 lines. Zero bloat, zero dependencies, zero vendor lock-in.

• Expressive: Everything you love—(Multi-)Agents, Workflow, RAG, and more.

• Agentic Coding: Let AI Agents (e.g., Cursor AI) build Agents—10x productivity boost!

Get started with Pocket Flow:

• To install, pip install pocketflowor just copy the source code (only 100 lines).
• To learn more, check out the video tutorial and documentation
• 🎉 Join our Discord to connect with other developers building with Pocket Flow!
• 🎉 Pocket Flow now has Typescript, Java, C++, Go, Rust and PHP versions!

Why Pocket Flow?

Current LLM frameworks are bloated... You only need 100 lines for LLM Framework!

	Abstraction	App-Specific Wrappers	Vendor-Specific Wrappers	Lines	Size
LangChain	Agent, Chain	Many (e.g., QA, Summarization)	Many (e.g., OpenAI, Pinecone, etc.)	405K	+166MB
CrewAI	Agent, Chain	Many (e.g., FileReadTool, SerperDevTool)	Many (e.g., OpenAI, Anthropic, Pinecone, etc.)	18K	+173MB
SmolAgent	Agent	Some (e.g., CodeAgent, VisitWebTool)	Some (e.g., DuckDuckGo, Hugging Face, etc.)	8K	+198MB
LangGraph	Agent, Graph	Some (e.g., Semantic Search)	Some (e.g., PostgresStore, SqliteSaver, etc.)	37K	+51MB
AutoGen	Agent	Some (e.g., Tool Agent, Chat Agent)	Many [Optional] (e.g., OpenAI, Pinecone, etc.)	7K (core-only)	+26MB (core-only)
PocketFlow	Graph	None	None	100	+56KB

How does Pocket Flow work?

The 100 lines capture the core abstraction of LLM frameworks: Graph!

From there, it's easy to implement popular design patterns like (Multi-)Agents, Workflow, RAG, etc.

✨ Below are basic tutorials:

Name	Difficulty	Description
Chat	☆☆☆ Dummy	A basic chat bot with conversation history
Structured Output	☆☆☆ Dummy	Extracting structured data from resumes by prompting
Workflow	☆☆☆ Dummy	A writing workflow that outlines, writes content, and applies styling
Agent	☆☆☆ Dummy	A research agent that can search the web and answer questions
RAG	☆☆☆ Dummy	A simple Retrieval-augmented Generation process
Batch	☆☆☆ Dummy	A batch processor that translates markdown into multiple languages
Streaming	☆☆☆ Dummy	A real-time LLM streaming demo with user interrupt capability
Chat Guardrail	☆☆☆ Dummy	A travel advisor chatbot that only processes travel-related queries
Majority Vote	☆☆☆ Dummy	Improve reasoning accuracy by aggregating multiple solution attempts
Map-Reduce	☆☆☆ Dummy	Batch resume qualification using map-reduce pattern
CLI HITL	☆☆☆ Dummy	A command-line joke generator with human-in-the-loop feedback
Multi-Agent	★☆☆ Beginner	A Taboo word game for async communication between 2 agents
Supervisor	★☆☆ Beginner	Research agent is getting unreliable... Let's build a supervision process
Parallel	★☆☆ Beginner	A parallel execution demo that shows 3x speedup
Parallel Flow	★☆☆ Beginner	A parallel image processing showing 8x speedup
Thinking	★☆☆ Beginner	Solve complex reasoning problems through Chain-of-Thought
Memory	★☆☆ Beginner	A chat bot with short-term and long-term memory
Text2SQL	★☆☆ Beginner	Convert natural language to SQL queries with an auto-debug loop
Code Generator	★☆☆ Beginner	Generate test cases, implement solutions, and iteratively improve code
MCP	★☆☆ Beginner	Agent using Model Context Protocol for numerical operations
A2A	★☆☆ Beginner	Agent wrapped with A2A protocol for inter-agent communication
Streamlit FSM	★☆☆ Beginner	Streamlit app with finite state machine for HITL image generation
FastAPI WebSocket	★☆☆ Beginner	Real-time chat interface with streaming LLM responses via WebSocket
FastAPI Background	★☆☆ Beginner	FastAPI app with background jobs and real-time progress via SSE
Voice Chat	★☆☆ Beginner	An interactive voice chat application with VAD, STT, LLM, and TTS.

👀 Want to see other tutorials for dummies? Create an issue!

How to Use Pocket Flow?

🚀 Through Agentic Coding—the fastest LLM App development paradigm-where humans design and agents code!

✨ Below are examples of more complex LLM Apps:

App Name	Difficulty	Topics	Human Design	Agent Code
Website Chatbot Turn your website into a 24/7 customer support genius	★★☆ Medium	Agent RAG	Design Doc	Flow Code
Danganronpa Simulator Forget the Turing test. Danganronpa, the ultimate AI experiment!	★★★ Advanced	Workflow Agent	Design Doc	Flow Code
Codebase Knowledge Builder Life's too short to stare at others' code in confusion	★★☆ Medium	Workflow	Design Doc	Flow Code
Build Cursor with Cursor We'll reach the singularity soon ...	★★★ Advanced	Agent	Design Doc	Flow Code
Ask AI Paul Graham Ask AI Paul Graham, in case you don't get in	★★☆ Medium	RAG Map Reduce TTS	Design Doc	Flow Code
Youtube Summarizer Explain YouTube Videos to you like you're 5	★☆☆ Beginner	Map Reduce	Design Doc	Flow Code
Cold Opener Generator Instant icebreakers that turn cold leads hot	★☆☆ Beginner	Map Reduce Web Search	Design Doc	Flow Code

• Want to learn Agentic Coding?

  • Check out my YouTube for video tutorial on how some apps above are made!

  • Want to build your own LLM App? Read this post! Start with this template!

---

🆕 Additional Feature: Universal LLM API Client

This fork includes an enhanced utils/llm_api.py module that provides a unified interface for multiple LLM providers:

Features

• Multiple Providers: OpenAI, Azure OpenAI, Anthropic, Google Gemini, DeepSeek, SiliconFlow, Local models, and Ollama
• Image Support: Attach images to prompts across compatible providers
• Easy Configuration: Simple environment variable setup with .env support
• Consistent Interface: Same API regardless of the underlying provider

Quick Start

from utils.llm_api import LLMClient

# Simple query with any provider
llm = LLMClient(provider="openai")
response = llm.query("What's the meaning of life?")

# Query with image (supported providers)
llm = LLMClient(provider="openai")
response = llm.query("Describe this image", image_path="photo.jpg")

# Backward compatibility with existing code
llm = LLMClient(provider="ollama")
messages = [{"role": "user", "content": "Hello!"}]
response = llm.call(messages=messages)

Supported Providers

Provider	Model	Notes
openai	GPT-4o	Supports images
anthropic	Claude 3.5 Sonnet	Supports images
gemini	Gemini 2.0 Flash	Supports images
azure	Azure OpenAI	Enterprise solution
deepseek	DeepSeek Chat	Cost-effective
siliconflow	DeepSeek R1	High performance
local	Self-hosted	Custom endpoint
ollama	Local models	Privacy-first

Environment Setup

Create a .env file with your API keys:

# OpenAI
OPENAI_API_KEY=your_openai_api_key

# Anthropic
ANTHROPIC_API_KEY=your_anthropic_api_key

# Google Gemini
GOOGLE_API_KEY=your_google_api_key

# Azure OpenAI
AZURE_OPENAI_API_KEY=your_azure_api_key
AZURE_OPENAI_MODEL_DEPLOYMENT=gpt-4o-ms

# DeepSeek
DEEPSEEK_API_KEY=your_deepseek_api_key

# SiliconFlow
SILICONFLOW_API_KEY=your_siliconflow_api_key

This makes it incredibly easy to switch between different LLM providers in your PocketFlow applications!